Defective pixel

Which clustering of defective pixels would obscure a subtle lesion ... 

Fig. 3.21. Defective pixels in a DR detector (a) mean thumbnail image (monochrome and colour deviation map) indicating no defective pixels and an almost perfect homogeneity.

Also typical for DR systems are defective pixel artifacts. These can occur in different forms. The... 

Fig. 4.9. Blurring at the top and bottom edges of the image due to crystallization (black arrows) of the a-Se detector (Agfa DM1000 DR). Also, clusters of defective pixels are visible (white arrows)...

General functions give access to the so-called camera datasheet, including information about the supplier, imager, its defect pixel as well as the intrinsic and extrinsic calibration of the camera. Furthermore, the camera operation mode and startup parameters can be set and operation status read. 

Required Data We must have enough data to calculate each of the figures of merit for which a pixel can be declared a defect. We usually have that data even if the customer does not have a Defective Pixel requirement. 

Algorithm Reporting defective pixels is strictly a logic check. For each requirement and pixel, does this pixel pass or fail Even though it may not be required. 

Definition There is no universal definition for a defective pixel. We include here a list of parameters normally considered, along with representative limits. 

Post Correction Nonuniformity. The PCNU is the standard deviation of the output after the nonuniformity correction process (see Section 10.4.4, PCNU). We are generally allowed to exclude some pixels from the standard deviation calculation - but any excluded pixels must be declared to be defective. That means that we have some latitude in the PCNU process - we may make PCNU better by excluding anomalous pixels - but at the cost of having more defective pixels. If the spec is very sfringent, we may need to devise an algorithm that finds the best compromise between PCNU and the number of defects. 

An additional step is needed - that of comparing the results from the two colors. In particular, the specification of defective pixels may require that we merge the defects from the two tests. This is primarily a counting operation, but it is also convenient to see the results in the form of a map. We will use the notation of the form A B to indicate the number of pixels that failed requirement A, but passed requirement B and A (or A) to indicate a pixel that passed (or failed) the requirements for color A, without regard to performance in color B. 

TABLE 104 Defective Pixel Count for a Two-Color Array... 

---